Elasticity of ion stuffing in chemically strengthened glass Adama Tandia, K. Deenamma Vargheese, John C. Mauro ⁎ Science and Technology Division, Corning Incorporated, Corning, New York 14831, USA abstract article info Article history: Received 8 March 2012 Received in revised form 16 April 2012 Available online 10 May 2012 Keywords: Ion exchange; Elasticity; Modeling The chemical strengthening of glass involves the stuffing of large ions into network sites previously occupied by smaller ions. Typically this involves an exchange of Li + or Na + ions in the glass for larger K + ions from a molten salt bath. This swapping of ions creates compressive stress in the surface of the glass, significantly in- creasing the strength of the final glass product. The magnitude of this compressive stress is governed by the linear network dilation coefficient (LNDC), which defines the amount of linear strain per unit of ion substitu- tion. However, the amount of strain attainable through ion exchange is much smaller compared to what is expected from as-melted versions of the same final glass composition. This effect, which we have termed the “network dilation anomaly,” is a result of the different local environment around the invading ion species in as-melted versus ion-exchanged glasses. A remaining question concerns the nature of the network strain due to ion stuffing. Using molecular dynamics simulations, we show that the strain induced by ion stuffing is entirely elastic. In other words, when a reverse ion exchange is performed to swap the original ions back into the glass, the initial volume of the as-melted glasses is entirely recovered. Moreover, we show that the local structural environment around the alkali ions is restored to the as-melted conditions. The elastic nature of ion stuffing demonstrates that the network dilation anomaly is not a result of plasticity, but rather a failure to achieve the full amount of expected elastic strain during ion exchange. The elasticity itself consists of both instantaneous and delayed contributions. © 2012 Elsevier B.V. All rights reserved. 1. Introduction The advent of ultra-thin chemically strengthened glass has revolu- tionized the personal electronics industry, enabling an entirely new class of touch-screen computers where glass is no longer simply a medium for visualizing information—it is now the primary interface through which the user interacts with the device. The thinness of the chemically strengthened cover glass enables high transparency, high touch sensitivity, and reduced weight for the final product. In addition, the dramatically improved strength of the glass makes it highly resis- tant to scratches and other damage during every day use [1]. The chemical strengthening of glass is achieved through an ion exchange process in which smaller alkali ions in the glass are replaced with larger alkali ions from a molten salt, a process originally pro- posed by Kistler [2] in 1962. The glass in question typically contains some concentration of a small alkali oxide, such as Li 2 O or Na 2 O. Upon immersion in a salt bath containing a larger alkali ion (such as K + ), the smaller alkali ions from the glass diffuse into the salt bath and are replaced with the larger ions from the salt. When these larger ions are stuffed into sites originally occupied by smaller ions in the glass, they create a compressive stress around the glass surface that greatly improves the strength and damage resistance of the final glass product. Through this ion stuffing, the chemical strengthening process is capable of achieving a significantly higher compressive stress compared to the traditional thermal tempering approach. Ion exchange is especially well suited for strengthening of thin glass, where thermal tempering is generally not a viable option [3–6]. The compressive stress achieved through the ion exchange pro- cess is given by σ z ðÞ¼ - BE 1-ν Cz ðÞ-C avg h i ; ð1Þ where E is the Young's modulus of the glass, ν is Poisson's ratio, and C(z) is the local concentration of the invading alkali ion at a penetra- tion depth z within the glass. The average concentration of the substituting cations, C avg , must be subtracted from C(z) so that the integral of σ(z) over the complete thickness of the glass is zero, i.e., to satisfy the force balance condition. Thus, the integral of the nega- tive (compressive) stress near the surface is exactly compensated by the integral of the positive (tensile) stress in the interior of the glass [7]. The parameter B in Eq. (1) is known as the linear network dilation coefficient (LNDC) [7–10] and can be considered as an analo- gous quantity to the thermal expansion coefficient. Whereas the thermal expansion coefficient defines the linear strain of a material Journal of Non-Crystalline Solids 358 (2012) 1569–1574 ⁎ Corresponding author. Tel.: + 1 607 974 2185; fax: + 1 607 974 2410. E-mail address: mauroj@corning.com (J.C. Mauro). 0022-3093/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.jnoncrysol.2012.04.021 Contents lists available at SciVerse ScienceDirect Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol